polygenic adaptation
Casto AM & Feldman MW 2011 Genome-wide association study SNPs in the Human Genome Diversity Project populations: does selection affect unlinked SNPs with shared trait associations? PLoS Genet 7:e1001226.
- given the polygenic nature of complex traits, selection may exert its influence on them by altering allele frequencies at many associated loci
- a possibility which has yet to be explored empirically
- empirical studies have found only a few examples of variants that have undergone "hard sweeps", where a favored allele arises by mutation and increases to high frequency in a population
- adaptation to new phenotypic optima may proceed through small allele frequency changes at many polymorphic loci
- selection on traits need not be limited to positive selection
- selection acts on traits not only to achieve new phenotypic optima, but also to maintain those optima once they have been reached, perhaps through balancing or negative selection
- there is solid evidence that these study groups represent instances of a polygenic response to selective pressure on an associated trait
- epistasis may be widespread among polymorphic sites in the human genome and can be difficult to detect, particularly as the number of putatively interacting loci considered increases
- epistasis has been frequently found to exist between loci in model organisms like flies and mice
- many, even most, GWAS SNPs are associated with more than one trait
- like epistasis, pleiotropy has been found to be common for variants influencing complex traits in model organisms
- we found multiple instances of it in the dataset that we studied here
- the presence of epistasis and pleiotropy can mean that the response of a variant to selection is dependent on the genetic background and the existence and strength of selection on other traits
- complexities we did not account for in our analysis
- recent work has proposed negative selection as one explanation for what appear to be common non-neutral patterns of variation in the human genome, such as the elevation of Fst values in genes and the reduction of diversity at sites linked to coding and regulatory regions
- our results suggest that both negative and positive selection may influence variation at GWAS SNPs
- their relative contributions, however, remain unclear